Search papers, labs, and topics across Lattice.
This paper benchmarks Vision-Language-Action (VLA) models on various hardware accelerators (GPUs, XPUs, NPUs) under Cost, Energy, and Time (CET) constraints to optimize on-robot deployment. They find that edge devices can be more cost and energy-efficient than high-end GPUs, and identify a two-phase inference pattern (compute-bound VLM, memory-bound Action Expert) that leads to hardware underutilization. Based on these insights, they propose DP-Cache and V-AEFusion, achieving significant speedups (up to 6x on NPUs) with minimal performance loss.
Edge NPUs can outperform flagship GPUs in cost and energy efficiency for on-robot VLA model deployment, but only with hardware-aware optimizations that tackle the models' distinct compute and memory-bound phases.
Vision-Language-Action (VLA) models are promising for generalist robot control, but on-robot deployment is bottlenecked by real-time inference under tight cost and energy budgets. Most prior evaluations rely on desktop-grade GPUs, obscuring the trade-offs and opportunities offered by heterogeneous edge accelerators (GPUs/XPUs/NPUs). We present a systematic analysis for low-cost VLA deployment via model-hardware co-characterization. First, we build a cross-accelerator leaderboard and evaluate model-hardware pairs under CET (Cost, Energy, Time), showing that right-sized edge devices can be more cost-/energy-efficient than flagship GPUs while meeting control-rate constraints. Second, using in-depth profiling, we uncover a consistent two-phase inference pattern: a compute-bound VLM backbone followed by a memory-bound Action Expert, which induces phase-dependent underutilization and hardware inefficiency. Finally, guided by these insights, we propose DP-Cache and V-AEFusion to reduce diffusion redundancy and enable asynchronous pipeline parallelism, achieving up to 2.9x speedup on GPUs and 6x on edge NPUs with only marginal success degradation. The example leaderboard website is available at: https://vla-leaderboard-01.vercel.app/.
